Power ShinyApp

The Power ShinyApp is a web-based interactive application developed using R’s Shiny framework. The developed app is designed to perform statistical power analysis for Two-Sample/Population Functional Hypothesis Test (Functional/Curve data) using multiple statistical methods. The application supports multiple data inputs, interactive visualisation, and statistical power computation to provide a flexible tool for simulation-based power analysis.

Quick Start (TL;DR)

# 1) Install R and RStudio (see detailed instructions below if you need help)

# 2) Get the code
# Option A: download ZIP from GitHub, then unzip and cd into the folder

# Option B: clone with Git
git clone https://github.com/mr-seydi/PowerShinyApp.git
cd PowerShinyApp

# 3) Open in RStudio and install packages
# In R console:
install.packages(c("shiny"))  # or use renv::restore() if renv.lock exists

# 4) Run the app

## From RStudio

# 1. Double-click `Power_ShinyApp.Rproj` to open the project.
# 2. In the Console, run: shiny::runApp()
# Shiny will automatically find `ui.R` and `server.R` in the project root.

## From Terminal

# Since this app has ui.R and server.R in the project root:
R -e "shiny::runApp('.', host='127.0.0.1', port=3838)"

Requirements

  • Operating System: Windows 10/11, macOS 12+ (Monterey) or newer, or a modern Linux distribution
  • R version: 4.2 or newer (older may work but is not supported)
  • RStudio: current desktop release
  • Internet: required on first run for package installation

Check your R version:

R.version.string

Install R

Windows

  1. Download R from CRAN: https://cran.r-project.org/bin/windows/base/

  2. Run the installer and accept the defaults unless you need a custom location.

  3. (Optional, CLI) If you use Chocolatey, open Admin PowerShell and run:

    choco install r.project -y

macOS

  1. Download the latest R pkg from CRAN: https://cran.r-project.org/bin/macosx/

  2. Double-click the .pkg and follow the prompts.

  3. (Optional, Homebrew):

    brew install --cask r

Linux

  • Ubuntu/Debian:

    sudo apt update
sudo apt install -y r-base

  • Fedora/RHEL/CentOS:

    sudo dnf install -y R

  • Arch:

    sudo pacman -S --noconfirm r

For system libraries needed by certain R packages (e.g., libcurl, openssl, xml2), see your distro docs.

Install RStudio

  • Download RStudio Desktop (Posit) for your OS: https://posit.co/download/rstudio-desktop/
  • Run the installer (macOS: drag app to Applications; Windows: follow the setup wizard; Linux: install the .deb/.rpm).
  • Launch RStudio.

Get the App Code

Option B — Download ZIP

  1. Go to https://github.com/mr-seydi/PowerShinyApp.
  2. Click Code ▸ Download ZIP.
  3. Unzip and open the folder.

Project structure (example):

PowerShinyApp/
├─ ui.R
├─ server.R
├─ Power_ShinyApp.Rproj
├─ R/                 # helper functions
├─ www/               # static assets (css/js/images)
├─ data/              # data files (if any)
├─ renv.lock          # pinned package versions (if using renv)
└─ README.md

Set Up R Packages

Open the project in RStudio (double-click Power_ShinyApp.Rproj) or set your working directory to the repo root:

setwd("/path/to/PowerShinyApp")

Run the App

From RStudio

  1. Double-click Power_ShinyApp.Rproj to open the project.

  2. In the Console, run:

    shiny::runApp()

    Shiny will automatically find ui.R and server.R in the project root.

From Terminal

Run Shiny on a specific host/port:

R -e "shiny::runApp('.', host='0.0.0.0', port=3838)"

Then open: http://localhost:3838

Introduction to Power ShinyApp

Power ShinyApp is your Tool for Statistical Power Analysis for Two-Sample/Population Functional Hypothesis Test. It provides a user-friendly platform for conducting advanced statistical power analyses of functional (curve) data. Whether you are in academia or industry, it simplifies complex calculations, enabling accurate, valid and efficient study planning, sample size estimation and data interpretation.

Dataset Selection

Choose how you want to create or upload mean functions (curves) for your two-group/population comparison:

  1. Baseline + Gaussian signal — Select a mean function (curve) of one group from six publicly available datasets and add a Gaussian signal to create the other group’s mean function.
  2. Two-group dataset — Select a dataset that represents the mean functions (curves) of two groups from various publicly available biomechanical datasets.
  3. Custom drawing — Draw custom mean functions (curves) directly on a coordinate plane.
  4. Pilot data upload — Upload your pilot data to perform power analysis.

Available Input Types

  1. Baseline Data
Pic2 Pic3
  1. Two-Sample Data

  1. Custom Curve
Pic5 Pic6
  1. Pilot Data

Pic7

Pic8

Pilot Data Requirements

Assume you have preliminary data gathered during a pilot study — a small-scale preliminary study to test the procedures and instruments of a larger, main study. This data helps researchers refine their methodology, estimate necessary sample sizes for the full study, identify potential problems, and assess the feasibility of their research.

The dataset you upload must have the following structure:

  • Columns: Each column should represent an individual curve.
  • Rows: Each row corresponds to an evaluation point, and there must be exactly 101 observation points (rows).

After uploading the dataset, you will select two groups of individual curves for analysis. The estimated smoothness (Noise FWHM) and the standard deviation of the noise functions will guide you in choosing the appropriate parameters for the next step.

If your raw data has a different number of evaluation points, you can resample to 101 evaluation points using the following R helper function (requires the fda package):

library(fda)

data_to_101evalpoints <- function(data){
  # data: matrix with rows = original time/eval points, columns = functions
  data_to_fd <- Data2fd(argvals = 0:(nrow(data) - 1), y = data)
  rangeval <- data_to_fd$basis$rangeval
  eval.points <- seq(rangeval[1], rangeval[2], length.out = 101)
  final_data <- eval.fd(fdobj = data_to_fd, evalarg = eval.points)
  return(final_data)
}

Parameters and Plots

Parameter Panels

This section allows you to define the characteristics of both the noise and, if applicable, the signal used in your analysis or simulation. Adjusting these parameters changes how the app generates or interprets your data, giving you full control over the input conditions.

Noise Parameters
  • Sample Size: number of functions (curves) in each group.
  • Mean (μ): average value of the noise.
  • Standard Deviation (σ): variability of the noise.
  • Full Width at Half Maximum (FWHM): controls the smoothness of the noise; higher FWHM means smoother noise.
Signal Parameters (shown only for baseline data or custom curves)
  • Center: location of the main effect or signal peak.
  • Signal FWHM (SFWHM): width of the signal.
  • Amplitude: height or strength of the signal.

Plots Section

Based on the parameters you choose, the app displays three plots side by side to help you visualize the data and check whether your settings make sense.

  • Noise Plot: shows one illustration of the simulated noise curve.
  • Signal Plot / Mean functions: shows the mean functions and the effect (difference between two mean functions) or signal (if using baseline data or custom curves).
  • Data Plot: shows one illustration of the generated sample functional data.

Visual examples

1. Baseline Data

Pic9 Pic10

2. Two-Sample Data

Pic11 Pic12

3. Custom Curve

Pic13 Pic14

4. Pilot Data

Pic15 Pic16

Power Calculation

  • Method: choose one or more statistical tests (IWT, TWT, SPM, F-max, ERL, IATSE). For more information, see the associated papers referred to in the paper.
  • Number of Iterations: how many simulations to run for estimating power (more iterations give more stable results but take longer).
  • Calculate Power: starts the power calculation.
  • Stop Calculation: interrupts long simulations.

The calculated statistical power is shown on the right side of this tab.

Notes & Tips

  • Use pilot data to estimate realistic noise parameters (FWHM and σ) whenever possible, this usually improves the accuracy of sample size recommendations.
  • When drawing custom curves, try to match the range and smoothness expected in your real measurements so simulations reflect realistic conditions.
  • Save commonly used parameter sets as presets for fast re-use during iterative study planning.
  • If you use Power ShinyApp in a publication, please cite the app and the methodology papers linked inside the app.
  • For bugs or feature requests, open an issue or contact the developers.

Troubleshooting

  • R not found / wrong version

    • Confirm in R console: R.version.string
    • On macOS, if using Homebrew, ensure /usr/local/bin or /opt/homebrew/bin precedes older R paths.

  • Packages fail to compile on Windows

  • Packages fail to compile on macOS

    • Install Xcode Command Line Tools: xcode-select --install

  • Linux missing system libs

    • Install libcurl, openssl, libxml2, harfbuzz, fribidi, freetype, etc., via your package manager.

  • Port already in use

    • Change port: shiny::runApp(port = 4242)

  • App can’t find files

    • Make sure your working directory is the project root. In RStudio: Session ▸ Set Working Directory ▸ To Project Directory.

  • White screen / nothing loads

    • Open the R console for errors; check the browser dev tools (Network/Console). Try options(shiny.fullstacktrace = TRUE).

How to Update

cd PowerShinyApp
git pull
# If using renv:
R -e "renv::restore()"

License

This project is licensed under the MIT License unless otherwise noted.

Citation

If you use this app in academic work, please cite this paper:

Sample Size Estimation for Functional Data Analysis with Web Application

Mohammad Reza Seydi, Johan Strandberg, Todd C. Pataky, and Lina Schelin

Acknowledgments

  • Shiny by Posit

Maintainer & Support

  • Maintainer: Reza Seydi
  • Email: mohammad.seydi@umu.se
  • Issues: Please open an issue here: https://github.com/mr-seydi/PowerShinyApp/issues